CN109102451A - A kind of anti-fake halftoning intelligent digital watermarking method of paper media's output - Google Patents
A kind of anti-fake halftoning intelligent digital watermarking method of paper media's output Download PDFInfo
- Publication number
- CN109102451A CN109102451A CN201810819980.6A CN201810819980A CN109102451A CN 109102451 A CN109102451 A CN 109102451A CN 201810819980 A CN201810819980 A CN 201810819980A CN 109102451 A CN109102451 A CN 109102451A
- Authority
- CN
- China
- Prior art keywords
- image
- watermark
- embedded
- halftoning
- fake
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0065—Extraction of an embedded watermark; Reliable detection
Abstract
The invention discloses a kind of anti-fake halftoning intelligent digital watermarking methods of paper media output, specific steps are as follows: host image is subjected to binary conversion treatment, binary conversion treatment is also carried out to the watermarking images of insertion simultaneously, form binary number sequence, as embeddable watermark encoder information, use random number seed as key, generate a series of two-dimensional random numbers pair, embedded location as insertion watermark encoder information, the watermark encoder information formed in S1 is embedded into the embedded location of screening, it is optimized using host image of the vision iterative algorithm to insertion watermark encoder information；Watermarking images are extracted using trained neural network model.The present invention can be produced with anti-fraud functional printing watermark, in copyright identification, anti-counterfeit package, all have wide practical use in terms of intelligent packaging, there is higher technology content, printing watermark embedded quantity is larger, will not significantly affect image perception quality, watermark has hiding, does not influence consumer and uses printed matter.
Description
Technical field
The present invention relates to the anti-fake halftoning intelligence numbers that a kind of digital watermarking production method more particularly to a kind of paper media export
Word watermarking method.
Background technique
Digital watermark technology can effectively improve the anti-infringement ability of digital media information.It is based on digital watermarking on the market at present
The packages printing anti-fake product of technology is also rarely found, and tracing it to its cause is because printing image and the imaging mechanism of digital picture have
The difference of essence.Printed matter is the tint hierarchy and color change that continuous tone is reproduced with the network point distribution of halftoning.Even
Continuous digital picture first has to carry out screening and color separation processing, the digital watermarking being embedded in originally is in this mistake during printing
It cannot be substantially detected by very havoc using scanning resampling quantization in journey.The present invention be directed to gray scales
The printing watermark manufacture that image is implemented, is not related to color image range.
The digital watermark technology towards half tone image also rarely has on domestic market mostly also in development both at home and abroad
Practical product occurs.Hagit Z. and Hel-Or proposed a kind of in the halftoning process of image in 2000, made
The halftoning digital image watermarking method of watermark information insertion is carried out with two different dither matrixs.Ming Sun Fu and
Oscar C.Au combined Halftone Algorithm for Error Diffusion in 2003, was embedded into watermarking images during halftone process
In image.Basic ideas are: generating the pseudo-random position of Information hiding using pseudo random number, modify halftone screen on these positions
Point is the coding for hiding watermark information.There is a collection of digital figure watermark algorithm to put forward based on this route, such as DHST, DHPT,
DHSPT, DHE and MDHED scheduling algorithm.Hereafter, Ming Sun Fu et al. has also been proposed core self-adjusting error diffusion algorithm, uses
Two error diffusion kernels rather than two different dither matrixs realize the embedding of watermark information in the halftoning process of image
Enter.There are also unit, such as Wuhan University, the research teachers and students such as Xi'an University of Technology to expand similar scientific research activity for the country.It is main
It is in progress are as follows: inquired into based on human visual system model, half color with watermarked information is minimized by progressive alternate
The collimation error changed the line map between picture and former consecutive image, to obtain the half tone image of high quality.
Digital watermarking can be generally divided into frequency domain insertion and be embedded in airspace.Frequency domain insertion is usually enterprising in continuous toned image
Capable, image is first transformed into frequency domain, then be embedded in watermark on frequency spectrum coefficient data, is then regenerated by inverse transformation
The advantages of digital picture, frequency domain is embedded in watermark is that have preferable holographic, can keep out image section defect and scanning to not
Neat influence of the problem to the extraction effect of watermark information.However the image screening for passing through printing process changes, the letter of frequency domain addition
Breath is largely destroyed, and is not easy that printing watermark is made.Method based on spatial transform, which just compares, to be suitble in screening process
It is embedded in watermark information, printing watermark is made.If being embedded in watermark bit letter over these locations using the isolated point randomly selected
Breath, it is small on being influenced on image vision, it is embeddable to contain much information, but the mistake that isolated dots are re-entered in the output for printing and scanning
It is difficult to be accurately positioned in journey, or even dot loss occurs, is not easy to detect the information of insertion.Another kind of spatial domain watermark scheme, according to
Select different screening modes that watermark is dissolved into the neighborhood mould of host's half tone image during screening according to watermark information
In formula, in the output for printing and after rescaning input, it is easier to detect watermark information.
In the link of detection watermark, existing detection algorithm is a reverse calculating process, introduces the print of formulation
Brush scan transformation function, by calculating the watermark information for deriving and being embedded in screening process, because of the printing and scanning of formulation
Model all cannot include complex conditions and random accidentalia in actual production process, so to the mapping mode of screening mode
Correct estimation cannot be made, very big deviation can be generally generated.The actually detected effect of watermark is undesirable.
Summary of the invention
The purpose of the present invention is to solve the above-mentioned problems, provides a kind of anti-fake halftoning intelligent digital of paper media's output
Watermarking method, it has the intelligent modeling that row data-driven is transformed into the dot of image undergone in printing and scanning, can
Intelligently to detect watermark information, while the advantages of put up a resistance to the behavior of potential secondary printing piracy.
To achieve the goals above, the present invention adopts the following technical scheme:
A kind of anti-fake halftoning intelligent digital watermarking method of paper media's output, specific steps are as follows:
S1: carrying out binary conversion treatment for host image, while also carrying out binary conversion treatment to the watermarking images of insertion, formed two into
The digital string of system, as embeddable watermark encoder information；
S2: using random number seed as key, generate a series of two-dimensional random numbers pair, as the embedding of insertion watermark encoder information
Enter position, the watermark encoder information formed in S1 is embedded into the embedded location of screening；
S3: it is optimized using host image of the vision iterative algorithm to insertion watermark encoder information, to keep embedded location
Watermark encoder Information invariability changes the network point distribution situation of neighborhood, make host image insertion watermark encoder information after with it is not embedding
Visual perception difference between image before entering watermark encoder information reaches small as far as possible；
S4: the gray level image of multiple 520 x520, training neural network, by a variety of Bayer moulds of used gray level image are used
Formula screening forms picture library, the gray level image of multiple 520 x520 is divided into the picture of 8x8, each picture corresponds to one kind and visits
Ear mold type meshing method prints out the gray level image of screening, and scanner using formatting, retrieves again
The bianry image of 520x520 forms network training data set, network instruction by obtained binary image segmentation at the small picture of 8x8
The labeled data for practicing data set is exactly the meshing method coding of each small pieces, and training neural network is used on these training datas
In the identification model of meshing method, the two-dimensional water mark bianry image of insertion is extracted using trained neural network model.
Before carrying out binary conversion treatment to watermarking images in the S1, reversible scramble first is carried out to watermarking images, increases and protects
Close property.
In the S1, when forming binary number sequence, using the mode of lossless compression-encoding, advantageously form few as far as possible
Binary number sequence.
In the S2, watermarking images are binary pictures, and the value of each pixel is 0 or 1 to change this binary bitmap
Watermark sequence value is sequentially embedded on the embedded location being randomly generated for a binary sequence.
In the S3, vision iterative algorithm does reference according to original consecutive image, explores in the neighborhood of each pixel several
Kind of pixel all carries out exchanged form, each iteration towards the direction that local mean square deviation reduces, and whole processes pixels are complete to be obtained
It is interim optimum results, then carries out next round and optimize point by point, until equal between half tone image and original continuous toned image
Until variance conjunction does not continue to decline.
In the S4 method particularly includes: 1) host image of watermarking images will be added by printer output；2) it exports
Image be scanned instrument resampling；3) to the image preprocessing of sampling, adjustment direction, scaled size, with threshold value appropriate
Binaryzation obtains and prints preceding 520x520 bianry image of the host image with watermark with same resolution ratio；4) it will obtain
520x520 bianry image be divided into the small pieces of 8x8 in order, each small pieces sequentially inputs trained neural network mould
Type, Neural Network model predictive go out the meshing method of each picture, and the two-dimensional water mark binary map being embedded in is assembled through image
Picture.
Binary conversion treatment is carried out to host image and watermarking images using Bayer shake screening algorithm in the S1, Bayer is trembled
Dynamic screening algorithm carries out binary conversion treatment to image using threshold matrix, and the image to screening is divided according to the size of dither matrix
Block matrix is rushed, carries out binaryzation by the threshold value of dither matrix setting in block matrix, the formula that Bayer dither matrix generates is as follows:
DnIt is n rank bayer matrix, 2nx2nDither matrix；UnIt is 2nx2nUnit matrix；Set D0=0；Repeatedly using formula (1)
In generation, calculates, and generates the Bayer dither matrix of n rank；Bayer dither matrix has good symmetry, is able to achieve preferable visual effect.
The watermarking images are embedded into the specific algorithm of host image are as follows:
1) different shake core D is selected1,D1', it is iterated to calculate in the way of formula (1), generates two 3 different rank Bayers
Tremble matrix D3And D3'；
2) entire host image line is divided into the fritter of 8x8, the corresponding watermark pixel value 0 or 1 of each fritter；
3) watermark pixel is embedded in host image and uses D if current wa pixel is 13The current 8x8 of dither matrix binaryzation is small
Block uses D if watermark pixel is 03' the current 8x8 fritter of dither matrix binaryzation.
Beneficial effects of the present invention:
It can be produced using the present invention with anti-fraud functional printing watermark, in copyright identification, anti-counterfeit package, intelligent packaging side
Face all has wide practical use；The present invention is fuzzy to printing scanning natural process mainly by the memory capability of neural network
Modeling, there is higher technology content；Printing watermark embedded quantity is larger, will not significantly affect image perception quality, and watermark has hidden
Hiding property does not influence consumer using printed matter, can identify the printing watermark of imitation, prevent watermark bootlegging, digital watermarking
The effect of Scanning Detction is good, it is complete to extract information, and have certain confidentiality.
Detailed description of the invention
Fig. 1 is the method flow diagram of this method；
Fig. 2 is watermark pictorial information；
Fig. 3 is host's picture used in embodiment one；
Fig. 4 is the image that one screening of embodiment is embedded in watermark information；
Fig. 5 is the image acquired after embodiment one is print scanned；
Fig. 6 is the watermark picture that embodiment one is extracted using the present invention；
Fig. 7 is host's picture used in embodiment two；
Fig. 8 is the image that two screening of embodiment is embedded in watermark information；
Fig. 9 is the image acquired after embodiment two is print scanned；
Figure 10 is the watermark picture that embodiment two is extracted using the present invention.
Specific embodiment
The invention will be further described with embodiment with reference to the accompanying drawing.
As shown in Figure 1, a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output, specific steps are as follows:
S1: carrying out binary conversion treatment for host image, while also carrying out binary conversion treatment to the watermarking images of insertion, formed two into
The digital string of system, as embeddable watermark encoder information；
Binary conversion treatment is carried out to host image and watermarking images using Bayer shake screening algorithm, Bayer shakes screening algorithm benefit
Binary conversion treatment is carried out to image with threshold matrix, the image to screening rushes block matrix according to the size division of dither matrix,
Binaryzation is carried out by the threshold value of dither matrix setting in block matrix, the formula that Bayer dither matrix generates is as follows:
DnIt is n rank bayer matrix, 2nx2nDither matrix；UnIt is 2nx2nUnit matrix；Set D0=0；Repeatedly using formula (1)
In generation, calculates, and generates the Bayer dither matrix of n rank；Bayer dither matrix has good symmetry, is able to achieve preferable visual effect.
S2: using random number seed as key, generate a series of two-dimensional random numbers pair, as insertion watermark encoder information
Embedded location, the watermark encoder information formed in S1 is embedded into the embedded location of screening, watermarking images are binary pictures,
The value of each pixel is 0 or 1 this binary bitmap to be changed into a binary sequence, in the embedded location being randomly generated
On, it is sequentially embedded watermark sequence value；
S3: it is optimized using host image of the vision iterative algorithm to insertion watermark encoder information, to keep embedded location
Watermark encoder Information invariability changes the network point distribution situation of neighborhood, make host image insertion watermark encoder information after with it is not embedding
Visual perception difference between image before entering watermark encoder information reaches small as far as possible；Vision iterative algorithm is according to original company
Continuous image does reference, explores several pixels to exchanged form in the neighborhood of each pixel, each iteration is all towards local mean square deviation
Reduced direction carries out, and whole processes pixels are complete, and obtain is interim optimum results, then carries out next round and optimizes point by point, directly
Until mean square deviation conjunction does not continue to decline between half tone image and original continuous toned image；
S4: the gray level image of multiple 520 x520, training neural network, by a variety of Bayer moulds of used gray level image are used
Formula screening forms picture library, the gray level image of multiple 520 x520 is divided into the picture of 8x8, each picture corresponds to one kind and visits
Ear mold type meshing method prints out the gray level image of screening, and scanner using formatting, retrieves again
The bianry image of 520x520 forms network training data set, network instruction by obtained binary image segmentation at the small picture of 8x8
The labeled data for practicing data set is exactly the meshing method coding of each small pieces, and training neural network is used on these training datas
In the identification model of meshing method, the two-dimensional water mark bianry image of insertion is extracted using trained neural network model.
Before carrying out binary conversion treatment to watermarking images in S1, reversible scramble first is carried out to watermarking images, increases confidentiality.
In S1, formed binary number sequence when, using the mode of lossless compression-encoding, advantageously form few as far as possible two into
The digital string of system.
In the S4 method particularly includes: 1) host image of watermarking images will be added by printer output；2) it exports
Image be scanned instrument resampling；3) to the image preprocessing of sampling, adjustment direction, scaled size, with threshold value appropriate
Binaryzation obtains and prints preceding 520x520 bianry image of the host image with watermark with same resolution ratio；4) it will obtain
520x520 bianry image be divided into the small pieces of 8x8 in order, each small pieces sequentially inputs trained neural network mould
Type, Neural Network model predictive go out the meshing method of each picture, and the two-dimensional water mark binary map being embedded in is assembled through image
Picture；
The picture library of training neural network model is established using the gray level image of 15 520x520；By the grayscale image of 15 520x520
As 24 kinds of bayer-pattern screenings, the picture library of a 24*15 is formd；By these image segmentations at the small picture of 8x8, each
Small picture corresponds to a kind of meshing method；Selection uses the preferable Bayer jitter mode of two kinds of experiment effects in the application, marks respectively
Note generates 2*15 kind Screening Image to this 15 gray level images for 0,1., these Screening Images are by printout, scanner weight
New sampling, format optimization, retrieves the bianry image of 520x520；Based on these bianry images after print scanned, point
The small pieces of 8x8 are cut, form network training data set, the labeled data of network training data set is exactly the screening side of each small pieces
Formula coding, the identification model of training neural network meshing method for identification on these training datasets；By trained mind
It saves through network into a model data, neural network is trained using 14 sub-pictures, verify nerve net with the 15th width figure
The effect of network identification model uses the 15th sub-picture as detection image, carries out watermark insertion experiment:
15th width image is once embedding according to the arrangement of watermark binary message using 0,1 two kinds of screening modes during screening
Enter watermark information；The halftone image that watermark is added passes through printer output, and the image of output is scanned instrument resampling, to adopting
The image preprocessing of sample, adjustment direction, scaled size are obtained with threshold binarization appropriate and are printed preceding half tune with watermark
The 520x520 bianry image of the same resolution ratio of image；The image is divided into the small pieces of 8x8 in order, each small pieces sequence is defeated
Enter trained neural network recognization model, neural network recognization model can predict the meshing method of each picture, figure
As the two-dimensional water mark bianry image being just embedded in after assembling.
The watermarking images are embedded into the specific algorithm of host image are as follows:
1) different shake core D is selected1,D1', it is iterated to calculate in the way of formula (1), generates two 3 different rank Bayers
Tremble matrix D3And D3'；
2) entire host image line is divided into the fritter of 8x8, the corresponding watermark pixel value 0 or 1 of each fritter；
3) watermark pixel is embedded in host image and uses D if current wa pixel is 13The current 8x8 of dither matrix binaryzation is small
Block uses D if watermark pixel is 03' the current 8x8 fritter of dither matrix binaryzation.
Embodiment one
The watermarking images of the present embodiment insertion are simple bianry image, and image is as shown in Fig. 2, this example describes basic stamping ink
The details implemented is printed, the host image used is a width gray level image, as shown in figure 3, gray level image Density Distribution is relatively uniform,
Watermark information sequence is directly embedded into the field battle array of the 8x8 of host image by the step of can saving selection watermark embedded location
In column, image printout result is had not significant impact, using the technology of the present invention, watermarking images, place are embedded on host image
Master image resolution ratio is 520x520, and insertion watermarking images are the 64x64 images an of binaryzation, and using 64 rows, 64 column 8x8 are small
The dot matrix meshing method of block carries out screening to host image, as a result the watermarking images of binaryzation is embedded into host image, such as
Shown in Fig. 4, the print scanned effect of image and the watermark information detected are as shown in Figure 5 and Figure 6, embedding information in the present embodiment
It measures larger, is a 64x64 1-bit information flow.
Embodiment two
The present embodiment is the application fields such as use is anti-fake in books, and package printed matter is anti-fake, and printing water is used in Books illustration
Copyright is published in print verifying, selects a width gray level image, as shown in fig. 7, verifying page through screening as one page in books, verifying
The printing copyright of version object, such printing watermark appearance looks elegant have dicoration than two dimensional code, and more safe and secret, equally
In high-grade packing box, the check card being beautifully printed can be made, includes printing watermark, as shown in figure 8, facilitating scanning
Verifying, concrete operations can be by the verification pages of books, and check card scanning input computer in package forms digital picture,
Recycle network by scan image as shown in figure 9, uploading to the verification page that manufacturer provides, selection commodity can be online
Examining product is certified products；As shown in Figure 10, the watermarking images information extracted using the present invention；The present invention is also needed and is serviced
Website is used cooperatively, if developing dedicated handheld scanning device will further improve service efficiency, two images screening is embedding
Enter watermark, print scanned image is used to train watermark detection model, the detection effect of scan image.
Embodiment three
The present invention also can be used in digital picture, be a kind of very distinctive digital watermarking effect, on the digital image
It is fine to detect watermark effect, is directly calculated using numerical value, speed is fast, is to be believed in Screening Image with the watermark that numerical method obtains
Breath.This method is difficult to copy, highly-safe, can solve digital watermarking to a certain extent, two dimensional code, pacifies present in bar code
Full problem, bigger advantage are that this digital watermarking is relatively beautiful, it is contemplated that have good user's attraction.
Above-mentioned, although the foregoing specific embodiments of the present invention is described with reference to the accompanying drawings, not protects model to the present invention
The limitation enclosed, those skilled in the art should understand that, based on the technical solutions of the present invention, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within protection scope of the present invention.
Claims (8)
1. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output, which is characterized in that specific steps are as follows:
S1: carrying out binary conversion treatment for host image, while also carrying out binary conversion treatment to the watermarking images of insertion, formed two into
The digital string of system, as embeddable watermark encoder information；
S2: using random number seed as key, generate a series of two-dimensional random numbers pair, as the embedding of insertion watermark encoder information
Enter position, the watermark encoder information formed in S1 is embedded into the embedded location of screening；
S3: it is optimized using host image of the vision iterative algorithm to insertion watermark encoder information, to keep embedded location
Watermark encoder Information invariability changes the network point distribution situation of neighborhood, make host image insertion watermark encoder information after with it is not embedding
Visual perception difference between image before entering watermark encoder information reaches small as far as possible；
S4: the gray level image of multiple 520 x520, training neural network, by a variety of Bayer moulds of used gray level image are used
Formula screening forms picture library, the gray level image of multiple 520 x520 is divided into the picture of 8x8, each picture corresponds to one kind and visits
Ear mold type meshing method prints out the gray level image of screening, and scanner using formatting, retrieves again
The bianry image of 520x520 forms network training data set, network instruction by obtained binary image segmentation at the small picture of 8x8
The labeled data for practicing data set is exactly the meshing method coding of each small pieces, and training neural network is used on these training datas
In the identification model of meshing method, the two-dimensional water mark bianry image of insertion is extracted using trained neural network model.
2. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In before carrying out binary conversion treatment to watermarking images in the S1, first to the reversible scramble of watermarking images progress, increase confidentiality.
3. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In, in the S1, formed binary number sequence when, using the mode of lossless compression-encoding, advantageously form few as far as possible two into
The digital string of system.
4. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In in the S2, watermarking images are binary pictures, and the value of each pixel is 0 or 1 this binary bitmap to be changed into one
A binary sequence is sequentially embedded watermark sequence value on the embedded location being randomly generated.
5. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In in the S3, vision iterative algorithm does reference according to original consecutive image, explores several pixels in the neighborhood of each pixel
To exchanged form, each iteration is all carried out towards the direction that local mean square deviation reduces, and whole processes pixels are complete, and obtain is the stage
Property optimum results, then carry out next round optimize point by point, until between half tone image and original continuous toned image mean square deviation close
Until not continuing to decline.
6. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In in the S4 method particularly includes: 1) host image of watermarking images will be added by printer output；2) image exported
It is scanned instrument resampling；3) to the image preprocessing of sampling, adjustment direction, scaled size, with threshold binarization appropriate,
It obtains and prints preceding 520x520 bianry image of the host image with watermark with same resolution ratio；4) it will obtain
520x520 bianry image is divided into the small pieces of 8x8 in order, each small pieces sequentially inputs trained neural network model,
Neural Network model predictive goes out the meshing method of each picture, and the two-dimensional water mark bianry image being embedded in is assembled through image.
7. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as described in claim 1, feature exist
In using Bayer shake screening algorithm to host image and watermarking images progress binary conversion treatment in the S1, Bayer shake adds
Net algorithm carries out binary conversion treatment to image using threshold matrix, and the image to screening rushes block according to the size division of dither matrix
Matrix carries out binaryzation by the threshold value of dither matrix setting in block matrix, and the formula that Bayer dither matrix generates is as follows:
DnIt is n rank bayer matrix, 2nx2nDither matrix；UnIt is 2nx2nUnit matrix；Set D0=0；Repeatedly using formula (1)
In generation, calculates, and generates the Bayer dither matrix of n rank；Bayer dither matrix has good symmetry, is able to achieve preferable visual effect.
8. a kind of anti-fake halftoning intelligent digital watermarking method of paper media's output as claimed in claim 7, feature exist
In the watermarking images are embedded into the specific algorithm of host image are as follows:
1) different shake core D is selected1,D1', it is iterated to calculate in the way of formula (1), generates two 3 different rank Bayers and tremble
Matrix D3And D3'；
2) entire host image line is divided into the fritter of 8x8, the corresponding watermark pixel value 0 or 1 of each fritter；
3) watermark pixel is embedded in host image and uses D if current wa pixel is 13The current 8x8 of dither matrix binaryzation is small
Block uses D if watermark pixel is 03' the current 8x8 fritter of dither matrix binaryzation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810819980.6A CN109102451B (en) | 2018-07-24 | 2018-07-24 | Anti-counterfeiting halftone intelligent digital watermark manufacturing method for paper media output |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810819980.6A CN109102451B (en) | 2018-07-24 | 2018-07-24 | Anti-counterfeiting halftone intelligent digital watermark manufacturing method for paper media output |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109102451A true CN109102451A (en) | 2018-12-28 |
| CN109102451B CN109102451B (en) | 2023-03-21 |
Family
ID=64847284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810819980.6A Active CN109102451B (en) | 2018-07-24 | 2018-07-24 | Anti-counterfeiting halftone intelligent digital watermark manufacturing method for paper media output |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109102451B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109948278A (en) * | 2019-03-29 | 2019-06-28 | 山东建筑大学 | Building Design aesthetic measure appraisal procedure and system based on width study |
| CN110189354A (en) * | 2019-04-18 | 2019-08-30 | 北京迈格威科技有限公司 | Image processing method and image processor, image processing equipment and medium |
| CN111161215A (en) * | 2019-12-09 | 2020-05-15 | 合肥迪宏自动化有限公司 | Detection apparatus capable of identifying color of printer |
| CN111611823A (en) * | 2020-05-26 | 2020-09-01 | 山东汇贸电子口岸有限公司 | Two-dimensional code identification optimization method |
| WO2020199500A1 (en) * | 2019-03-31 | 2020-10-08 | 深圳大学 | Clone-proof qr code and anti-counterfeiting qr code authentication method for qr code |
| CN112488902A (en) * | 2020-12-21 | 2021-03-12 | 重庆紫光华山智安科技有限公司 | Optimization method, system, medium and terminal for self-adaptive hidden watermark |
| CN113705806A (en) * | 2020-05-22 | 2021-11-26 | 中国人民银行数字货币研究所 | Neural network training method, image encoding and decoding method, device and medium |
| CN114139665A (en) * | 2021-10-14 | 2022-03-04 | 珠海丹德图像技术有限公司 | Anti-counterfeiting embedding method and identification method of negotiable securities, anti-counterfeiting device and negotiable securities |
| CN115499556A (en) * | 2022-09-19 | 2022-12-20 | 浙江工业大学 | Digital printing screening method based on machine learning iteration |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030095683A1 (en) * | 2001-08-28 | 2003-05-22 | Kayvan Najarian | Automated digital watermarking methods using neural networks |
| WO2012022086A1 (en) * | 2010-08-16 | 2012-02-23 | 中兴通讯股份有限公司 | Electronic entrance ticket process method and device thereof |
| CN104616244A (en) * | 2015-01-24 | 2015-05-13 | 河南师范大学 | Image watermark embedding and extracting method based on back propagation (BP) neural network compressed domain |
| CN106096668A (en) * | 2016-08-18 | 2016-11-09 | 携程计算机技术（上海）有限公司 | The recognition methods of watermarked image and the system of identification |
-
2018
- 2018-07-24 CN CN201810819980.6A patent/CN109102451B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030095683A1 (en) * | 2001-08-28 | 2003-05-22 | Kayvan Najarian | Automated digital watermarking methods using neural networks |
| WO2012022086A1 (en) * | 2010-08-16 | 2012-02-23 | 中兴通讯股份有限公司 | Electronic entrance ticket process method and device thereof |
| CN104616244A (en) * | 2015-01-24 | 2015-05-13 | 河南师范大学 | Image watermark embedding and extracting method based on back propagation (BP) neural network compressed domain |
| CN106096668A (en) * | 2016-08-18 | 2016-11-09 | 携程计算机技术（上海）有限公司 | The recognition methods of watermarked image and the system of identification |
Non-Patent Citations (2)
| Title |
|---|
| 刘文涛等: "基于半色调数字图像的抗打印的水印算法研究", 《齐鲁工业大学学报》 * |
| 徐兴等: "基于半色调数字图像的水印技术", 《齐鲁工业大学学报》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109948278A (en) * | 2019-03-29 | 2019-06-28 | 山东建筑大学 | Building Design aesthetic measure appraisal procedure and system based on width study |
| WO2020199500A1 (en) * | 2019-03-31 | 2020-10-08 | 深圳大学 | Clone-proof qr code and anti-counterfeiting qr code authentication method for qr code |
| CN110189354A (en) * | 2019-04-18 | 2019-08-30 | 北京迈格威科技有限公司 | Image processing method and image processor, image processing equipment and medium |
| CN110189354B (en) * | 2019-04-18 | 2021-12-28 | 北京迈格威科技有限公司 | Image processing method, image processor, image processing apparatus, and medium |
| CN111161215B (en) * | 2019-12-09 | 2023-03-31 | 合肥中科迪宏自动化有限公司 | Detection apparatus capable of identifying color of printer |
| CN111161215A (en) * | 2019-12-09 | 2020-05-15 | 合肥迪宏自动化有限公司 | Detection apparatus capable of identifying color of printer |
| CN113705806B (en) * | 2020-05-22 | 2024-01-02 | 中国人民银行数字货币研究所 | Network training, encoding and decoding method, device and medium for digital watermarking |
| CN113705806A (en) * | 2020-05-22 | 2021-11-26 | 中国人民银行数字货币研究所 | Neural network training method, image encoding and decoding method, device and medium |
| CN111611823A (en) * | 2020-05-26 | 2020-09-01 | 山东汇贸电子口岸有限公司 | Two-dimensional code identification optimization method |
| CN112488902A (en) * | 2020-12-21 | 2021-03-12 | 重庆紫光华山智安科技有限公司 | Optimization method, system, medium and terminal for self-adaptive hidden watermark |
| CN112488902B (en) * | 2020-12-21 | 2024-02-06 | 重庆紫光华山智安科技有限公司 | Optimization method, system, medium and terminal for self-adaptive hidden watermark |
| CN114139665A (en) * | 2021-10-14 | 2022-03-04 | 珠海丹德图像技术有限公司 | Anti-counterfeiting embedding method and identification method of negotiable securities, anti-counterfeiting device and negotiable securities |
| CN115499556A (en) * | 2022-09-19 | 2022-12-20 | 浙江工业大学 | Digital printing screening method based on machine learning iteration |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109102451B (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109102451A (en) | A kind of anti-fake halftoning intelligent digital watermarking method of paper media's output | |
| US7599099B2 (en) | Image processing apparatus and image processing method | |
| JP4000316B2 (en) | Generation of figure codes by halftoning using embedded figure coding | |
| CN103886545B (en) | Transformation domain printing-scanning resistant digital watermarking method based on computer-generated holography | |
| CN102254295A (en) | Color halftoning image watermarking algorithm based on support vector machine | |
| Amiri et al. | Robust watermarking against print and scan attack through efficient modeling algorithm | |
| CN101122995A (en) | Binary image digit water mark embedding, extraction method and device | |
| CN109815653A (en) | A kind of extraction of pdf Text Watermarking and comparison method based on deep learning | |
| CN113052745B (en) | Digital watermark model training method, ceramic watermark image manufacturing method and ceramic | |
| CN107578362A (en) | It is embedded in the method for secret information in the picture based on textures synthesis | |
| CN113158583A (en) | End-to-end text image watermark model establishing method based on deep learning | |
| US7586627B2 (en) | Method and system for optimizing print-scan simulations | |
| CN113222804B (en) | Ceramic process-oriented up-sampling ceramic watermark model training method and embedding method | |
| CN112217958B (en) | Method for preprocessing digital watermark carrier image irrelevant to device color space | |
| Son et al. | Watermark detection from clustered halftone dots via learned dictionary | |
| CN109671011A (en) | Embedding grammar, extracting method and the electronic equipment of copyright information | |
| CN101751655B (en) | Methods for implanting and extracting digital watermark and device thereof | |
| Cu et al. | Hiding security feature into text content for securing documents using generated font | |
| CN113065407B (en) | Financial bill seal erasing method based on attention mechanism and generation countermeasure network | |
| Cu et al. | A robust data hiding scheme using generated content for securing genuine documents | |
| Abedini et al. | Structure-aware halftoning using the iterative method controlling the dot placement | |
| JP3684181B2 (en) | Image processing apparatus and image processing method | |
| CN111127483B (en) | Color picture processing method, device, equipment, storage medium and system | |
| JP2003179761A (en) | Adaptive prediction coding and decoding method, device thereof and recording medium with adaptive prediction coding and decoding program recorded thereon | |
| JP3684178B2 (en) | Image processing apparatus and image processing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |